¹INSERM U365, Institut Curie, Section Recherche 26, rue d'Ulm, 75 248 - Paris Cedex 05, France

²CNRS-UMR 147, Institut Curie, Section Recherche 26, rue d'Ulm, 75 248 - Paris Cedex 05, France

³INSERM U509, Institut Curie, Section Recherche 26, rue d'Ulm, 75 248 - Paris Cedex 05, France

⁴MD, Department of Pathology, Institut Curie, Section Médicale, 26 rue d'Ulm, 75 248 - Paris Cedex 05 - France

Correspondence to: J Wietzerbin, E-mail: Jeanne.Wietzerbin@curie.fr

Ewing sarcoma is the second most common bone tumor in childhood. Despite aggressive chemotherapy and radiotherapy strategies, the prognosis of patients with metastatic disease remains poor. We have recently reported that Ewing tumor cell proliferation was strongly inhibited by IFN- and to a lesser degree by IFN-. Moreover, under IFN- treatment, some cell lines undergo apoptosis. Since the possibility of using IFNs for Ewing tumor treatments may be of interest, we have evaluated the efficacy of Hu-IFNs in a nude mice model of Ewing tumor xenografts. The results reported here show that human type I IFNs, Hu-IFN- and Hu-IFN- impaired tumor xenograft take and displayed an anti-growth effect toward established xenografts. Furthermore, we have also shown that combined therapy with Hu-IFNs and ifosfamide (IFO), an alkylating agent widely used in high-dose chemotherapy of Ewing tumors, results in a strong antitumor effect. Pathological analysis showed that Hu-IFN-/IFO and Hu-IFN-/IFO were characterized by a dramatic decrease in the mitotic index and marked necrosis, as well as extensive fibrosis associated with numerous calcifications. To our knowledge, this is the first demonstration of a potential antitumor effect of human type I IFNs and IFO on Ewing tumors, providing a rational foundation for a promising therapeutic approach to Ewing sarcoma.

Oncogene (2002) 21, 7700-7709. doi:10.1038/sj.onc.1205881

Ewing's sarcoma; IFN-; IFN-; ifosfamide; xenograft

Introduction

Ewing's sarcoma (ET) is a primitive neuroectodermal tumor characterized by a specific chromosomal translocation t(11 : 22) (q24:q12) (Aurias et al., 1983; Horowitz et al., 1997) which results in the production of the EWS-Fli-1 fusion protein, a chimeric transcription factor shown to display oncogenic properties (Delattre et al., 1992; Bailly et al., 1993; May et al., 1993; Ouchida et al., 1995; Kovar et al., 1996). Ewing's sarcoma is the second most common bone tumor in childhood. Advances in therapy including the use of multiple chemotherapy and radiotherapy have led to a survival rate of 50% patients after 5 years. Despite aggressive treatment strategies such as high-dose chemotherapy with stem cell rescue in combination with surgery and/or extended radiotherapy, the prognosis for patients with metastatic disease remains poor (Wexler et al., 1996; Fizazi et al., 1998; Rosito et al., 1999; Kushner and Meyers, 2001; Paulussen et al., 2001; Pinkerton et al., 2001) Therefore, more effective new therapeutic strategies are required to improve the outcome and survival of these patients.

In recent years, a lot of information has accumulated concerning potential interest in the use of cytokines for cancer therapy (Gutterman, 1994). Among cytokines, interferons (IFNs) are of particular interest due to their pivotal physiological role in the control of cell proliferation (Tanneberger and Harelia, 1996; Grander et al., 1997). Many studies have shown that Hu-IFNs are able to inhibit tumor cell proliferation followed in some cases by an apoptotic response. Hu-IFNs have been shown to antagonize the cell proliferation induced by various growth factors. Hu-IFNs are also immunostimulatory molecules which induce MHC expression leading to an increase in cytotoxic T lymphocyte activity, enhance the generation of T helper cells, activate NK cells and induce tumoricidal macrophage activity (Gutterman, 1994; Tanneberger and Harelia, 1996; Grander et al., 1997; Borden, 1998). Thus, these immunomodulatory activities probably contribute to an Hu-IFN-mediated antitumor response.

Type I Hu-IFNs, Hu-IFN- and Hu-IFN- are widely used at high dose with significant success, in adult patients presenting chronic hepatitis B or C (Rumi et al., 1997; Chapman et al., 2001), or with multiple sclerosis, for which recombinant-Hu-IFN- is a good recombinant therapeutic drug for the treatment of relapsing remitting multiple sclerosis (Hall et al., 1997; Ross et al., 2000; Comi et al., 2001). Type I Hu-IFNs have also been shown to be active in several human hematological malignancies such as chronic myelogenous leukemia (Borden, 1998; Talpaz, 2001), myeloma (Grander, 2000), hairy cell leukemia (Kurzrock et al., 1991; Gutterman, 1994) and lymphoma (Wollina et al., 2001). Studies of the effects of Hu-IFNs in non-hematopoietic tumors are ongoing. Preclinical and clinical studies on melanoma, colorectal carcinoma and prostate cancer suggest that type I Hu-IFNs may constitute a promising biotherapeutic agent for these tumors (Dong et al., 1999; Borden et al., 2000; Ozawa et al., 2001).

Clinical data from patients receiving Hu-IFN therapy indicated a correlation between the in vitro sensitivity of malignant cells to the direct antiproliferative effect of Hu-IFNs, and the in vivo clinical effects (Strander and Einhorn, 1996). Few clinical trials have sought to incorporate Hu-IFN into combination chemotherapy regimens against solid tumors. Various studies comparing therapy with antitumor chemotherapeutic agents versus the same treatment with Hu-IFN, demonstrated an increased objective response rate with the addition of Hu-IFN compared with cytotoxic agents alone (Wadler and Schwartz, 1990; Wadler and Schwartz, 1997; Faderl et al., 1999; Daponte et al., 2000; Todesco et al., 2000).

We have recently reported a detailed in vitro study of the effect of type I Hu-IFNs, Hu-IFN- and Hu-IFN- on the proliferation of tumoral cells derived from Ewing's sarcoma. Our results showed that Ewing's tumor cells were responsive to the antiproliferative effects of Hu-IFN- and to a lesser degree of Hu-IFN-. Moreover, under Hu-IFN- treatment, some Ewing's tumor cells undergo apoptosis (Sancéau et al., 2000). These observations are in line with published data describing the presence of specific functional receptors for type I Hu-IFNs in series of pediatric tumors including Ewing's tumors and in cell lines derived from these tumors (Rosolen et al., 1997).

Although many studies have examined and sometimes demonstrated the usefulness of Hu-IFNs for the treatment of patients with cancer, no data are available for Ewing's tumors. Since the possibility of using Hu-IFNs for Ewing's tumor treatment may be of interest, we evaluated the efficacy of Hu-IFNs in a relevant in vivo nude mouse model of Ewing's tumors. The results reported here show that type I Hu-IFNs, Hu-IFN- and Hu-IFN- impaired engrafting of Ewing's tumors in nude mice and displayed an antigrowth effect toward established xenografts. Ifosfamide (IFO) is a DNA-alkylating agent largely used in combined modalities for the therapy of different types of cancer such as small cell lung cancer and sarcoma (Sandler, 1998; Nemati et al., 2000). IFO is also widely used in high-dose combined chemotherapy of Ewing's tumors (Wexler et al., 1996). The results reported here clearly show a strong synergistic inhibition of Ewing's tumor xenograft growth by combined therapy with Hu-IFNs and IFO. Taken together, our findings provide a rational foundation for a promising therapeutic approach to Ewing's sarcoma.

Results

Effect of type I Hu-IFNs on EW-7-Ewing xenograft tumor take

Since xenografts are good preclinical models for evaluating the potential antitumor effects of Hu-IFNs (Brosjö et al., 1985; Horton et al., 1999), we decided to investigate the effect of Hu-IFN- and Hu-IFN- on Ewing's tumor xenografts. In the first experiment, we evaluated the efficacy of human type I Hu-IFNs, Hu-IFN- and Hu-IFN- in reducing tumor take of EW-7-Ewing's tumor by starting the treatment 3 days after tumor implantation. Control mice (n=12) received daily intratumoral injections of diluent. Hu-IFNs were administered at a dose of 8´10⁵ units daily, from Monday to Friday, for 3 weeks (n=12 for each Hu-IFN) (see protocol 1).

EW-7-tumor takes were significantly inhibited in the mice treated with Hu-IFN- as well as with Hu-IFN- as shown in Figure 1. Both Hu-IFNs impaired tumor growth compared with the linear tumor growth observed in control mice. Inhibition of tumor growth was observed as early as 12 days after the start of Hu-IFN treatment. While tumors in control mice reached a mean volume of 100 mm³ (range, 63-258 mm³), tumor volume in the treated groups did not exceed 25 mm³, i.e. respectively 0-63 mm³ for Hu-IFN- (P<0.004) and 0-63 mm³ for Hu-IFN- (P<0.001). At treatment stoppage, on day 25, more than half of treated mice did not present palpable tumors (Hu-IFN- six out of 12 mice, Hu-IFN- eight out of 12 mice), indicating a clear suppressive effect of Hu-IFNs on EW-7-Ewing's tumor engrafting.

Upon withdrawal of Hu-IFNs, tumor growth resumed in three mice treated with Hu-IFN-, and in four mice treated with Hu-IFN-, while six out of 12 Hu-IFN--treated mice and eight out of 12 Hu-IFN--treated mice remained free of tumors after 54 days.

Effect of Hu-IFN- and - on EW-7-Ewing tumor growth

We next initiated a second set of experiments where the sensitivity of EW-7-Ewing's tumor to Hu-IFN- and Hu-IFN- was studied in established tumors. Thus, mice bearing tumors of volume 100-300 mm³ were subjected to Hu-IFN treatments (8´10⁵ units daily, from Monday to Friday) as described for protocol 2. As shown in Figure 2, the mean relative tumor volume (RTV) increase linearly and rapidly in control mice between day 6 and day 33. Most control mice were ethically sacrificed on day 33 when the tumor grafts reached a volume of 3500 mm³. The time-course of tumor progression was significantly decreased after Hu-IFN treatments. Hu-IFN- as well as Hu-IFN- led to a substantial growth delay of 12 days (P<0.02) and 19 days (P<0.0002), respectively, compared to 8 days in the case of control mice. The response to Hu-IFN- was greater than that to Hu-IFN-. Administration of Hu-IFN- led to a TGI of 70% and a 2.4-fold increase in the TGDI, while Hu-IFN- treatment led to a TGI of 50% with 1.5-fold increase in the TGDI. Similar results were obtained in three independent experiments. Similar data were noted following subcutaneous administration of IFNs. We performed a typical experiment where IFN- was injected subcutaneously at two different doses in mice carrying established tumors (1´10⁵ and 8´10⁵ units per mouse). The dose response antigrowth effect obtained by subcutaneous administration was indeed similar to that observed by injecting interferon intratumorally. Subcutaneous administration of 8´10⁵ units of IFN- leads to a TGI of 73% and a 2.2-fold increase in the TGDI, while treatment with 1´10⁵ units leads to a TGI of 39% and a 1.6-fold increase in the TGDI (Figure not shown).

Response of EW-7-Ewing tumor xenografts to Hu-IFN- or Hu-IFN- treatment combined with ifosfamide treatment

It was shown recently that addition of IFO to Ewing's tumor standard chemotherapy improves the outcome and survival of patients (Wexler et al., 1996; Rosito et al., 1999; Kushner and Meyers, et al., 2001; Pinkerton et al., 2001). Thus, a combined therapy using IFO with Hu-IFN was tested. Preliminary experiments showed that 120 mg/kg IFO maximally (100%) inhibited growth of established EW-7-xenografts while a significant effect of 66% inhibition was observed at a dose of 60 mg/kg. Hu-IFN- or Hu-IFN- was administered at 1´10⁵ units/day, a dose defined in previous experiments as giving a 20% growth inhibition.

Two parallel studies of combinations were carried out in EW-7-tumor-bearing mice, given either the Hu-IFN-/IFO or the Hu-IFN-/IFO combination. IFO (60 mg/kg) was injected (i.p.) on days 1, 2 and 3. Intratumoral injections of Hu-IFN- or Hu-IFN- (1´10⁵ units/mouse, daily from Monday to Friday) started at day 2, and were continued for 3 weeks. Each study included a control group of mice and groups treated with Hu-IFNs or IFO alone. Figure 3a,b shows the curves obtained from the mean RTV data for each group. Almost linear tumor growth was observed in control mice. If we considered the twelfth day of treatment, when the first control mice was sacrificed, the average RTV was reduced from 5.2 (range 4.5-7.3) in control mice to 4 (range 1.6-5.7, P<0.02) in Hu-IFN- mice, and to 3.7 (range 2.1-5.3, P<0.02) in Hu-IFN- mice (Figure 3). The tumor growth was substantially delayed after IFO treatment, with the RTV reduced either to 1.6 (range 0.6-2.8, P<0.0003, Figure 3a) or to 2.5 (range 1.5-7.1, P<0.02, Figure 3b) compared with the RTV for control mice. The combination of Hu-IFN with IFO resulted in a dramatic decrease in tumor growth whatever the Hu-IFN combined (Hu-IFN- as well as Hu-IFN-), with a mean RTV of 0.87 (range 0.18-1.8, P<0.0001) for the Hu-IFN-/IFO combination, and 1.2 (range 0.05-2.5, P<0.0006) for the Hu-IFN-/IFO combination.

All control mice were ethically sacrificed by days 18-20 (Figure 3a,b). Treatments with Hu-IFNs were prolonged up to day 40 for the surviving mice, i.e. Hu-IFN-- or Hu-IFN--treated groups, as well as Hu-IFNs/IFO combinations. At this time, Hu-IFN-/IFO and Hu-IFN-/IFO treatments led to respective TGDs of 34 and 41 days compared with 16-21 days for treatment with IFO alone (depending on the experiments). For the control mice and Hu-IFN-treated mice, the growth delay was respectively 9 and 12 days. These results demonstrated a strong inhibition of EW-7-tumor xenograft growth by treatment combining human type I Hu-IFNs and IFO.

Response of COH-Ewing tumor xenografts to Hu-IFN- or Hu-IFN- treatment combined with ifosfamide treatment

To test the efficacy of Hu-IFNs/IFO combination on another Ewing's tumor type, we performed an experiment with COH-tumor xenograft. This cell line derived from a metastatic tumor resistant to current protocols using a combination of local surgery and radiation with systemic chemotherapy. Although the optimal IFO dose was higher than the dose used for EW-7-tumor xenografts, the combined treatment Hu-IFNs/IFO gave a very strong tumor growth inhibition. Figure 4 shows the curves obtained from the mean RTV data of each group. A rapid almost linear tumor growth was observed in control mice. If we considered the ninth day of treatment, when the first control mice were sacrificed, the average RTV was reduced from 7.6 (range 5.4-11) in control mice to 6 (range 5.05-9.18, P<0.02) in Hu-IFN- mice and 4.4 (range 3.43-5.76, P<0.01) in Hu-IFN- mice (Figure 4). After IFO treatment, the tumor growth was delayed with the RTV reduced to 4.9 (range 0.51-12.64, P<0.1) compared with the RTV for control mice. The combination of Hu-IFNs with IFO resulted in a consequent decrease in tumor growth whatever the Hu-IFN combined with a mean RTV of 1.1 (range 0.55-1.9, P<0.002) for the Hu-IFN-/IFO combination, and 1 (range 0.51-1.55, P<0.001) for the Hu-IFN-/IFO combination. These results suggest a common mechanism of action of the Hu-IFNs/IFO combination independently of Ewing's tumor type.

Pathological analysis of EW-7-Ewing tumor xenograft

Groups of eight mice were injected with diluent (control), IFO, Hu-IFN- or Hu-IFN- alone or combined with preinjection of IFO. All mice were sacrificed on day 15 of treatment, and tumor fragments were fixed in formalin. For pathological analysis, mitotic and apoptotic indexes, extent of necrosis and of fibrosis, and the presence of calcifications were assessed (Freneaux et al., 2000). This analysis showed that the groups of tumors treated by Hu-IFN-/IFO or Hu-IFN-/IFO were characterized by a dramatic decrease in the mitotic index and marked necrosis. These tumors were also characterized by extensive fibrosis, associated with numerous calcifications (Figure 5 and Table 1). A slight increase in apoptosis index was also observed after Hu-IFN-/IFO treatment, compared to control or to IFO or Hu-IFNs treatment alone.

Activation of Stat-1 in response to Hu-IFN- orHu-IFN- treatment alone or combined with IFO inEW-7-tumor xenografts

To evaluate whether treatment of mice with IFNs alone or combined with IFO results in the activation of signaling molecules in Ewing tumor xenografts, we analysed the state of activation of Stat-1 (i.e. Ser/Tyr phosphorylated Stat-1) in the tumor samples used for pathological analysis (Figure 5 and Table 1). Extracts of tumor fragments obtained from mice (n=4) injected with diluent (Control), IFO, Hu-IFN- or Hu-IFN- alone or combined with preinjection of IFO were analysed by Western blot. As shown in Figure 6, phosphorylated Stat-1 was not detectable in control tumor extracts whereas a very low but detectable level of phosphorylated Tyr701-Stat-1 protein was observed in the tumor extracts from IFO treated mice. As expected, IFN- or IFN- treatments induced the activation of Stat-1 with an apparent pSer727-Stat-1 protein. Interestingly, combined IFN-/IFO or IFN-/IFO treatment enhanced the level of pS727/pTyr701 phosphorylated Stat-1 protein which is known to be required for maximal activation of transcription by Stat-1 (Horvath and Darnell, 1997).

Discussion

The data presented here show that type I Hu-IFNs impaired Ewing's tumor xenograft engrafting, and provoked tumor growth delay of already established tumors in nude mice. Although in vitro studies have suggested that Ewing's tumor-derived cell lines were more sensitive to the antiproliferative effect of Hu-IFN- than of Hu-IFN- (Sancéau et al., 2000), both Hu-IFNs were equally effective in impairing tumor growth when administered just after tumor engrafting. It can be hypothesized that, in addition to their inhibition of cell proliferation through direct action on the tumor cells, Hu-IFNs might inhibit the production and action of autocrine growth factors and angiogenic stimulating factors (such as bFGF, or VEGF), responsible for the generation of new blood vessels (Dong et al., 1999; Slaton et al., 1999; Borden et al., 2000).

The antitumor response induced by Hu-IFN- seems superior to that of Hu-IFN- as shown by the higher tumor growth delay upon administration of Hu-IFN- compared to Hu-IFN- administration. These results are in agreement with our previous observations showing a more powerful antiproliferative effect of Hu-IFN- than Hu-IFN- in Ewing's tumor-derived cell lines (Sancéau et al., 2000). Despite the fact that all human type I IFNs bind to the same receptor and activate a common set of signaling factors, there is now accumulating evidence of differences in their signaling pathways. In addition, several reports suggest that Hu-IFN- may have greater efficacy in solid tumors (Pfeffer et al., 1998; Runkel et al., 1998; Stark et al., 1998). In our studies, growth inhibition of tumors is likely due to the direct antiproliferative activity of human type I IFNs. This is supported by the fact that the Hu-IFNs used in this study are of human origin and do not cross-react appreciably with the host mouse cells (Biron, 2001). However, it cannot be excluded that human IFNs acting on tumors might, via the induction of soluble factors, indirectly promote the activity of immune cells such as NK cells and macrophages that may enhance the anti-tumor response (Mizumo and Yoshida, 1998).

It is crucial to improve therapy of Ewing's tumor in advanced-stage disease, in older children and adults, with already micrometastatic disease at diagnosis. Less than 20% of these patients survive despite aggressive therapy consisting of surgery and/or radiation and chemotherapy (Fizazi et al., 1998; Rosito et al., 1999; Kushner and Meyers, 2001; Paulussen et al., 2001; Pinkerton et al., 2001). IFO is an alkylating agent widely used in high-dose combined chemotherapy, and inclusion of IFO in standard chemotherapeutic regimens was reported to provide benefits for Ewing's sarcoma patients (Wexler et al., 1996; Fizazi et al., 1998; Rosito et al., 1999; Kushner and Meyers, 2001; Paulussen et al., 2001; Pinkerton et al., 2001). Various in vitro and in vivo studies indicated that Hu-IFNs potentiate the activity of clinically useful drugs in a variety of human tumors and tumor models (Borden, 1998; Qin et al., 1998; Eck et al., 2001; Tada et al., 2001). A recent report describes a biochemotherapeutic approach in which Hu-IFN- was used as an adjuvant to chemotherapy with some success in relapsed melanoma and carcinomas (Wadler and Schwartz, 1990; Borden, 1998; Daponte et al., 2000).

The present study shows that Hu-IFN treatment may be advantageously associated with IFO administration. The combination was highly effective in suppressing tumor growth in nude mice bearing Ewing's tumor xenografts. From several separate experiments, it appears that tumor growth delays achieved with the Hu-IFN-/IFO or Hu-IFN-/IFO combination were repeatedly superior to those induced by a single agent, showing a potentiation of their efficacy. It should, however, be pointed out that the Hu-IFN-/IFO combination resulted in a greater tumor growth delay. With regard to the state of phosphorylation of Stat-1, a major transcription factor involved in IFN-regulated gene expression (Horvath and Darnell, 1997; Ihle, 2001), combined treatment with IFNs and IFO results in a marked increase in Serine/Tyrosine phosphorylation of Stat-1, compared to that induced by IFNs treatment alone. This may explain, at least in part, the synergistic inhibitory effect observed between IFNs and IFO on tumor growth.

Under the conditions examined here, the two type I Hu-IFNs exhibited synergy with IFO against tumoral progression. The Hu-IFN/IFO combination is effective in inhibiting the growth of xenografts derived from primary (EW-7) as well as metastatic (COH) tumors. The mechanism of interaction of IFO and Hu-IFNs remains to be understood. The action of the alkylating agent and that of Hu-IFNs could be independent. An alternative hypothesis is that Hu-IFNs enhance IFO cytotoxicity by increasing the proportion of tumor cells in G1-S phase, thus making them more susceptible to the cell cycle-specific actions of IFO. Alkylating agents such as IFO, able to act on non-dividing cells, make these drugs particularly attractive for the treatment of tumors because of their potential capacity to kill quiescent tumor cells while residual disease is minimal. IFO requires intracellular metabolism to be active. Its efficacy is dependent on different CYP450 which are in turn modulated by essentially 3A4 and 2B6 (Monshower et al., 1996). One possibility could be that Hu-IFNs modulate enzymes responsible for IFO degradation; decrease in these enzymes could increase the bioavailability of IFO. Hu-IFN-induced inhibition of CYP450 has been described (Monshower et al., 1996; Brockmeyer et al., 1998).

The studies presented in this report provide a rationale for carrying on with further investigations into the mechanism of interaction between the antiproliferative cytokines-Hu-IFNs - and a chemotherapeutic drug such as IFO. In children and young adults (<16 years), few data on the use of Hu-IFNs in other pathologies have been published to date (Strander and Einhorn, 1996; Tanneberger and Harelia, 1996; Wörle et al., 1999; Garmendia et al., 2001). The optimal timing and dosage for administration of Hu-IFNs alone or in combination with IFO remain to be evaluated.

Interferon/IFO treatment may provide an alternative approach to the treatment of patients with Ewing's tumors that are resistant to conventional chemotherapy and in the treatment of patients with minimal residual disease.

Materials and methods

Mice

Six- to seven-week-old Swiss nu/nu female mice (Iffa-Credo-France) maintained under specified pathogen-free conditions were used. Their care and housing were in accordance with the institutional guidelines of the French Ministerial Ethics Committee (Ministère de l'Agriculture et de la Pêche, Direction de la Santé et de la Protection Animale, Paris, France) and under the supervision of authorized investigators.

Tumor cells

Two cell lines derived from Ewing's sarcoma were used in this study: (1) wild-type p53 EW-7 cells, primary tumor localized on the scapula, (2) wild-type p53 COH cells, metastatic tumor localized on femur (Kovar et al., 1993; Hamelin et al., 1994). They were established as transplantable tumors by subcutaneous (s.c.) injection of 20´10⁶ cells. Xenografts were maintained in vivo by sequential passaging of s.c. implants with an engraftment success rate exceeding 90%.

Reagents

Recombinant human-IFN- 2b (Bioferon^Ò) was a gift from BioSidus (Buenos Aires, Argentina). Recombinant human-IFN- 1a (REBIF^Ò) was a gift from Ares Serono (Switzerland). Ifosfamide (Holoxan) was from Asta Medica Laboratories (France). Monoclonal antibodies against pS727-STAT-1, pY701-STAT-1 were from Upstate Biotechnology.

Formulation and administration

Interferons were diluted in PBS-BSA (NaCl 140 mM, Na₂HPO₄ 4.5 mM, KCl 2.5 mM, KH₂PO₄ 1 mM, containing 200 g/ml of purified bovine serum albumin, fraction V, -Sigma-) (diluent). Ifosfamide was given intraperitoneally. When given combined, drugs were injected separately. Mice in the control groups received diluent.

In vivo experimental design

Mice received 10-20 mm³ subcutaneous grafts of tumor fragments. When tumors reached a minimal volume of 60 mm³, mice were individually identified and randomly assigned to the control or treated groups (eight to 12 mice per group), and the treatments were started. Mice were sacrificed before the tumor volume reached 3500 mm³. Each mouse received the graft of one tumor piece s.c. over the flank through a 5 mm incision, and 40 to 50 mice were similarly transplanted in each experiment. Only mice with a reliable tumor growth were selected for the trial i.e. tumor volume more than 60 mm³ within 2-3 weeks after transplantation.

Tumor growth and evaluation of antitumoral effect

Tumor volumes were evaluated twice a week by measuring two perpendicular diameters with calipers. Tumor volume (V) was calculated using the following equation:

where a is the width of the tumor (small diameter), and b the length (large diameter), both in millimeters.

The individual relative tumor volume (RTV) was defined as Vx/V1, where Vx is the volume in mm³ at a given time and V1 at the start of treatment. Mean RTV and standard deviation were calculated for each group. Drug efficacy was expressed as the percentage tumor growth inhibition (% TGI), calculated using the equation 100-(T/C´100), where T is the mean RTV of the treated tumor and C is the mean RTV in the control group at the time of sacrifice of the first mouse in the control group. Tumor growth delay (TGD) was calculated as the time required for the tumor volume to increase fourfold over the initial volume at the start of treatment. Tumor growth delay index (TGDI) was calculated as the mean treated/control TGD ratio.

Protocol 1

Hu-IFN- or Hu-IFN- was injected into the side of EW-7-tumor implantation 3 days after tumor graft, according to the following regimens: group (a) diluent 0.1 ml; group (b) Hu-IFN- (8´10⁵ Units/0.1 ml/mouse); group (c) Hu-IFN- (8´10⁵ Units/0.1 ml/mouse). Hu-IFNs injections were repeated for 5 days every week (Monday to Friday) for 3 weeks.

Protocol 2

Diluent or Hu-IFNs (8´10⁵ Units/0.1 ml/mouse) was injected daily into established EW-7-tumors (5 days per week) for 5 weeks, according to the following regimens: group (a) diluent 0.1 ml; group (b) Hu-IFN-; group (c) Hu-IFN-.

Protocol 3. Combination of ifosfamide and Hu-IFNs

Ifosfamide at 60 mg/kg was administered (i.p.) to nude mice carrying an established EW-7-Ewing's tumor for the first 3 days, at the start of treatment. The second day, 1´10⁵ Units/mouse of Hu-IFN- or Hu-IFN- were injected into the tumor concomitantly with i.p IFO, according to the following regimens: group (a) diluent 0.1 ml; group (b) IFO 60 mg/kg; group (c) Hu-IFN-; group (d) Hu-IFN-; group (e) Hu-IFN-/IFO; group (f) Hu-IFN-/IFO. Hu-IFNs injections were continued five times a week for 3 weeks.

A similar protocol was performed with nude mice carrying an established COH-Ewing's tumor, except that the IFO dose was administered at 120 mg/kg (i.p.).

Protocol 4-programmed for pathological analysis

Groups of mice, carrying an established EW-7-tumor, received injection of diluent (control), IFO, Hu-IFN- or Hu-IFN- alone or combined with IFO injection, according to the schedule described as protocol 3. All mice were sacrificed on day 15 of treatment. Tumors were measured and fragments were fixed in formalin for microscopic analysis. Tissue sections (4 ) were stained by hematoxylin-eosin and analysed histologically. Five histological criteria were assessed: proliferation index, apoptotic index, extent of necrosis and of fibrosis, and presence of microcalcifications. For the evaluation of proliferation and apoptotic indexes, a representative field of the tumor was chosen by examination at low power. This representative field was then analysed at high power (´400) and the numbers of mitoses and of apoptotic bodies were recorded. Fibrosis and calcifications were assessed as absent (0), not extensive (1) or extensive (2). The extent of necrosis (%) was related to the surface of the tissue specimen analysed.

Western blot analysis

Total proteins from fresh tumor tissue (20-25 mg) were extracted in T-PER^Ô Tissue Extraction Reagent (150 l-Pierce-PerBio Science, France). After centrifugation (5 min, 10 000 r.p.m. at 4°C), total homogenates were diluated in 500 l of lysis buffer (50 mM HEPES pH 7.6, 150 mM NaCl, 1 mM EDTA, 1.5 mM MgCl₂, antiprotease/antiphosphatase mixture (aprotinin, leupeptin, NaF, ortho-vanadate, -glycerophosphate) containing 1% (w/v) n-octyl--D-glucoside (Surin et al., 2002). After lysis, the total extracts were centrifugated for 10 min at 10 000 r.p.m. at 4°C. Aliquots of the supernatants (50 g protein) were submitted to Western blot analysis as previously described (Sancéau et al., 2000).

Statistical analysis

The treated and control groups were compared using a one-sided Student's t-test. All P-values are one-sided with statistical significance defined as P<0.05.

Acknowledgements

The authors would like to thank BioSidus (Buenos Aires, Argentina) for the gift of human Hu-IFN--2b (Bioferon^Ò), and Ares Serono (Basel, Switzerland) for the gift of human Hu-IFN- 1a (REBIF^Ò). We thank Professor Gilbert Lenoir (Dept Pédiatrie, Hôpital Necker-Enfants-Malades, Paris, France) for providing the Ewing cells used in this study. We are grateful to Vincent Bordier, Claire Chevalier and Christophe Alberti for technical assistance for animal experimentation, and Catherine Silvestri for cell culture. This work was supported by grants from the Institut National Scientifique et de la Recherche Medicale (INSERM), the Association pour le Recherche Contre le Cancer (ARC) and BioSidus (Argentina).

Aurias A, Rimbaut C, Buffe C, Dubousset J, Mazabraud A. (1983). N. Engl. J. Med., 309: 496-497.

Bailly RA, Bosselut R, Zucman J, Cormier F, Delattre O, Roussel M, Thomas G, Gysdael J. (1993). Mol. Cell. Biol., 14: 3230-3241.

Biron CA. (2001). Immunity, 14: 661-664. Article MEDLINE

Borden EC. (1998). Oncologist, 3: 198-203. MEDLINE

Borden EC, Lindner D, Dreicer R, Hussein M, Peereboom D. (2000). Cancer Biol., 10: 125-144.

Brockmeyer NH, Barthel B, Mertins L, Goos M. (1998). Chemotherapy, 44: 174-180.

Brosjö O, Bauer HCF, Broström LA, Nilsonne U, Nilsson OS, Reinholt FP, Strander H, Tribukait B. (1985). Cancer Res., 45: 5598-5602.

Chapman BA, Stace NH, Edgar CL, Bartlett SE, Frampton CM, Scahill SL, Jennings LC. (2001). N. Z. Med. J., 114: 103-104.

Comi G, Filippi M, Barkhof F, Durelli L, Edan G, Fernandez O, Hartung H, Seeldrayers P, Sorensen PS, Rovaris M, Martelli V, Hommes OR. (2001). Lancet., 19: 1576-1582.

Daponte A, Ascierto PA, Gravina A, Lelucci MT, Palmieri G, Comella P, Cellerino R, DeLena M, Marini G, Comella G. (2000). Cancer, 89: 2630-2636.

Delattre O, Zucman J, Plougastel B, Desmaze C, Melot T, Peter M, Kovar H, Joubert I, de Jong P, Rouleau G, Aurias A, Thomas G. (1992). Nature, 359: 162-165. MEDLINE

Dong Z, Greene G, Pettaway C, Dinney CPN, Eue I, Iu W, Bucana CD, Balbay MD, Bielenberg D, Fidler IJ. (1999). Cancer Res., 59: 872-879. MEDLINE

Eck S, Alavi JB, Judy K, Philips P, Alavi A, Hackney D, Cross P, Hughes J, Gao G-p, Wilson JM, Propert K. (2001). Human Gene. Therapy, 12: 97-113.

Faderl S, Kantarjian HM, Talpaz M. (1999). Oncology, 13: 169-180. MEDLINE

Fizazi K, Dohollou N, Blay JY, Guerin S, Le Cesne A, Andre F, Pouillard P, Tursz T, Nguyen BB. (1998). J. Clin. Oncol., 16: 3736-3743. MEDLINE

Freneaux P, Stoppa-Lyonnet D, Mouret E, Kambouchner M, Nicolas A, Zafrani B, Vincent-Salomon A, Fourquet A, Magdelenat H, Sastre-Garau X. (2000). Br. J. Cancer, 83: 1318-1322.

Garmendia G, Miranda N, Borroso S, Longchong M, Martinez E, Ferrero J, Porrero P, Lopez-Saura P. (2001). J. Interferon Cytokine Res., 21: 31-38.

Grander D, Sangfelt O, Erickson S. (1997). Eur. J. Haematol., 59: 129-135. MEDLINE

Grander D. (2000). Acta. Oncol., 39: 801-805.

Gutterman JU. (1994). Proc. Natl. Acad. Sci. USA, 91: 1198-1205. MEDLINE

Hall GL, Compston A, Scolding NJ. (1997). Trends Neurosci., 20: 63-67.

Hamelin R, Zucman J, Melot T, Delattre O, Thomas G. (1994). Int. J. Cancer, 57: 336-340. MEDLINE

Horowitz ME, Malawer MM, Woo SY, Hicks MJ. (1997). Principles and Practice of Pediatric oncology 3rd edn. Pizzo PA and Poplack DG (eds) Philadelphia: Lippincott-Raven Publishers, pp. 831-863.

Horton HM, Hernandez P, Parker E, Barbhart KM. (1999). Cancer Res., 59: 4064-4068. MEDLINE

Horvath CM, Darnell JrJE. (1997). Curr. Opin. Cell. Biol., 9: 233-239. Article MEDLINE

Kovar H, Auinger A, Jug G, Aryee D, Zoubek A, Salver-Kuntschik M, Gadner H. (1993). Oncogene, 8: 2683-2690. MEDLINE

Kovar H, Aryee DN, Jud G, Henockl JG, Schemper M, Delattre O, Thomas G, Gadner H. (1996). Cell. Growth Differ., 7: 429-437. MEDLINE

Kurzrock R, Talpaz M, Gutterman JU. (1991). Br. J. Haematol., 79: 17-20.

Kushner B, Meyers PA. (2001). J. Clin. Oncol., 19: 870-880. MEDLINE

May WA, Lessnick SL, Braun BS, Klemsz M, Lewis BC, Lunsford LB, Hromas R, Denny CT. (1993). Mol. Cell. Biol., 13: 7393-7398. MEDLINE

Ihle JN. (2001). Curr. Opin. Cell. Biol., 13: 211-217. Article MEDLINE

Mizumo M, Yoshida J. (1998). Cancer Immunol. Immunother., 47: 227-232. MEDLINE

Monshower M, Witkamp RF, Nujmeijer SM, Van Amsterdam JG, Van Miert AS. (1996). Toxicol. Appl. Pharmacol., 137: 237-244.

Nemati F, Livartowski A, De Cremeux P, Bourgeois Y, Arvelo F, Pouillard P, Poupon MF. (2000). Clin. Cancer Res., 6: 2075-2086.

Ouchida M, Ohno T, Fujimura Y, Rao VN, Reddy ESP. (1995). Oncogene, 11: 1049-1054. MEDLINE

Ozawa S, Shinohara H, Kanayama HO, Bruns CJ, Bucana CD, Ellis LM, Davis DW, Fidler IJ. (2001). Neoplasia, 3: 154-164.

Paulussen M, Ahrens S, Dunst J, Winkelmann W, Exner GU, Kotz R, Amann G, Dockorn-Dworniczak B, Harms D, Müller-Weihreich S, Welte K, Kornhuleer B, Janka-Schaub G, Göbel U, Treuner J, Voûte PA, Zoubek A, Gadner H, Jürgens H. (2001). J. Clin. Oncol., 6: 1818-1829.

Pfeffer LM, Dinarello CA, Herberman RB, Williams BRG, Borden EC, Bordens R, Walter MR, Nagabhushan TL, Trotta PP, Pestka S. (1998). Cancer Res., 58: 2489-2499. MEDLINE

Pinkerton CR, Bataillard A, Guillo S, Obeerlin O, Fervers B, Philip T. (2001). Eur. J. Cancer, 37: 1338-1344.

Qin XQ, Tao N, Dergay A, Moy P, Fawell S, Davis A, Wilson JM, Barsoum J. (1998). Proc. Natl. Acad. Sci., USA, 95: 14411-14416. MEDLINE

Rosito P, Mancini AF, Rondelli R, Abate ME, Pession A, Bedei L, Bacci G, Picci P, Mercuri M, Ruggieri P, Frezza G, Campanacci M, Paolucci G. (1999). Cancer, 86: 421-428.

Rosolen A, Todesco A, Colamonici OR, Basso G, Frascella E. (1997). Modern Pathology, 10: 55-56.

Ross C, Clemmesen KM, Svenson M, Sorensen PS, Koch-Henriksen N, Skovgaard GL, Bendzen K and the Danish Multiple Sclerosis Study Group. (2000). Ann. Neurol., 48: 706-712. Article MEDLINE

Rumi MG, Santagostino E, Morfini M, Gringeri A, Tagariello G, Chistolini A, Pontisso P, Tagger A, Colombo M, Mannucci PM and the Hepatitis Study Group of the Association of Italian Hemophilia Centers. (1997). Blood, 89: 3529-3533.

Runkel L, Pfeffeer L, Lewerenz M, Monneron D, Yang CH, Murti A, Pellegrini S, Goelz S, Uze G, Mogensen K. (1998). J. Biol. Chem., 273: 8003-8008. MEDLINE

Sancéau J, Hiscott J, Delattre O, Wietzerbin J. (2000). Oncogene, 18: 3372-3383.

Sandler AB. (1998). Semin. Oncol., 25: 38-41.

Slaton JW, Perrotte P, Inoue K, Dinney CPN, Fidler IJ. (1999). Clin. Cancer Res., 5: 2726-2734. MEDLINE

Stark GR, Kerr IM, Williams BR, Silvermann RH, Schreiber RD. (1998). Annu. Rev. Biochem., 67: 227-264. Article MEDLINE

Strander H, Einhorn S. (1996). Biotherapy, 8: 213-218. MEDLINE

Surin B, Rouillard D, Bauvois B. (2002). Int. J. Mol. Med., 10: 25-31.

Tada H, Maron DJ, Choi EA, Barsoum J, Lei H, Xie Q, Liu W, Ellis L, Moscioni D, Tazelaar J, Fawell S, Qin X, Propert KJ, Davis A, Fraker DL, Wilson JM, Spitz FR. (2001). J. Clin. Invest., 108: 83-95.

Talpaz M. (2001). Semin. Hematol., 38: 22-27.

Tanneberger S, Harelia P. (1996). J. Interferon Cytokine Res., 16: 339-346. MEDLINE

Todesco A, Carli M, Iacona I, Ninfo V, Rosolen A. (2000). Cancer, 89: 2661-2666.

Wadler S, Schwartz EL. (1990). Cancer Res., 50: 3473-3486. MEDLINE

Wadler S, Schwartz EL. (1997). The Oncologist, 2: 254-267.

Wexler LH, DeLaney TF, Tsokos M, Avila N, Steinberg SM, Weaver-McClure L, Jacobson J, Jarosinski P, Hijazi YM, Balis FM, Horowitz ME. (1996). Cancer, 78: 901-911.

Wollina U, Looks A, Meyer J, Knopf B, Koch HJ, Liebold K, Hipler UC. (2001). J. Amer. Acad. Dermatol., 44: 253-260.

Wörle H, Maass E, Köhler B, Treuner J. (1999). Eur. J. Pediatr., 158: 344-345.

Figure 1 Effect of type I interferons on EW-7-tumor xenograft tumor take. Mice (12 mice per group) received injections in the side of tumor implantation with Hu-IFN- or Hu-IFN- (8´10⁵ units/mouse), starting 3 days after tumor graft, for 5 days every week. (A) Mean of tumor volume (mm³) n=12. Bars indicate s.e. (B) Individual tumor volume (mm³) at the end of the experiment (day 25)

Figure 2 Effect of Hu-IFN- and Hu-IFN- on established EW-7-tumor xenograft growth. Tumor-bearing mice were injected intratumorally, 5 days every week for 5 weeks. The graph represents the mean RTV of 12 mice per group at different times. Bars represent s.e. Relative tumor volume (RTV) and s.e. were calculated as indicated in Materials and methods

Figure 3 Strong inhibition of EW-7-tumor xenograft growth by combined Hu-IFN-/IFO (A) or Hu-IFN-/IFO (B) treatments. IFO at 60 mg/kg was administered (i.p.) to tumor-bearing mice for the first 3 days at the start of treatment. The second day, 1´10⁵ units/mouse of Hu-IFN- or Hu-IFN- were injected intratumorally concomitantly with i.p. IFN injections were continued five times a week for 3 weeks

Figure 4 Inhibition of COH-tumor xenograft growth by combined Hu-IFN-/IFO or Hu-IFN-/IFO treatments. IFO at 120 mg/kg was administered (i.p.) to COH-tumor-bearing mice for the first 3 days at the start of treatment. The second day, 1´10⁵ units/mouse of Hu-IFN- or Hu-IFN- were injected intratumorally concomitantly with i.p. IFO. IFN injections were continued five times a week

Figure 5 Pathological analysis of EW-7-tumor xenografts from mice treated by Hu-IFN-/IFO or Hu-IFN-/IFO combinations. Tumor-bearing mice (eight mice per group) received injection of diluent, IFO, Hu-IFN- or Hu-IFN- alone or combined with IFO, according to the schedule described in Figure 3. All mice were sacrificed on day 15 of treatment. Tumor fragments were fixed in formalin and tissue sections (4 ) were stained by hematoxylin-eosin. (a) Control: small cell tumors showing numerous mitotic features (open arrow) and few apoptotic bodies (solid arrows) (´250). (b) Ifosfamide treatment: numerous apoptotic bodies (solid arrows and few mitoses (solid arrows) (´250). (c) IFN-/Ifosfamide treatment-C1: extensive fibrosis (fi) with microcalcification (mc) (´120). C2: large dystrophic tumor cells, no mitosis and numerous apoptotic bodies (solid arrows) (´250). (d) IFN-/Ifosfamide treatment: D1: extensive fibrosis (fi) with numerous apoptotic bodies (solid arrows) (´250). D2: few and dystrophic cells scattered in abundant fibrosis (´250)

Figure 6 Phosphorylation state of Stat-1 proteins in EW-7-tumor xenografts from mice treated with Hu-IFN- or Hu-IFN- alone or combined with IFO. Protein extracts from fresh tumor materials used for pathological analysis of EW-7 tumor xenografts described in Figure 5 were used. Total lysates obtained from our four mice, representative of the eight mice of each group (Cont, IFO, IFN-, IFN-, IFO+IFN- and IFO+IFN-) were subjected to SDS-PAGE and Western blot analysis. Membranes were probed with antibodies directed against the activated phosphorylated Stat-1 proteins, i.e. pS727-STAT-1 and pY701-STAT-1

Table 1 Pathological analysis of tumors treated by IFN-/IFO and IFN-/IFO